Identifikasi Jenis Penyakit Daun pada Tanaman Pisang Menggunakan Metode Faster Region Convolutional Neural Network (Faster R-CNN) Berbasis Android

Abstract

Banana plants are widely recognized by the Indonesian community and are easy to cultivate in specialized or freely grown land. However, the productive growth of banana plants in Indonesia is decreasing due to several factors that result in a waste of banana harvest, leading to negative impacts on the quantity and quality of the fruit. Currently, the use of technology is still insufficient for rapid plant disease identification. In the field of informatics, digital image processing can be utilized to efficiently and effectively handle plant diseases for growers. In this study, the Faster Region Convolutional Neural Network (Faster R-CNN) method is used to identify leaf diseases in banana plants, consisting of three types of diseases: Yellow Sigatoka, Black Sigatoka, and Cordana. A total of 600 data points were used, with 480 data points for training and 120 data points for testing. The use of the Faster Region Convolutional Neural Network (Faster R-CNN) method for identifying leaf diseases in banana plants achieved an accuracy rate of 91.66%. Based on the obtained accuracy value, it can be concluded that the system built using the Faster Region Convolutional Neural Network (Faster R-CNN) method is effective in identifying leaf diseases in banana plants.